With the vigorous development of Internet of things, IoT gateways as the communication bridge of the whole IoT system has attracted more and more people’s attention. So the objective of this project is to measure and evaluate the performance of two kinds of IoT gateways including Arduino Uno and Vinnter IoT gateway. In order to reach this objective, the concrete implementation is to create a demonstrator via JAVA program to establish serial communication and measure the latency and throughput of respective IoT gateways. The measurement is implemented under a specific scenario which is to control the action of an actuator such as a LED connecting with IoT gateways via GPIO pins. Finally, the results will be presented via statistical data and line charts. Through analyzing the results, the performance of these two IoT gateways can be compared in terms of latency and throughput easily. Besides implementing the measurement in practice, this paper also involves the scientific research of many different industrial-level IoT gateways and their performance are compared on both hardware and software sides. Consequently, this thesis will show you the comparison of various IoT gateways both in theory and in practice.

The schedule for the jobs in a real-time system can have a huge impact on how the system behave. Since real-time systems are common in safety applications it is important that the scheduling is done in a valid way. Furthermore, one can enhance the performance of the applications by minimizing data latency and jitter. A challenge is that jobs in real-time systems usually have complex constraints making it too time consuming to minimize data latency and jitter to optimality. The purpose of this report is to investigate the possibility of creating high quality schedules using heuristics, with the goal to keep the computational time under one minute. This will be done by comparing three different algorithms that will be used on real scheduling instances provided by the company Arcticus. The first algorithm is a greedy heuristic, the second one a local search and the third one is a metaheuristic, simulated annealing. The results indicate that the data latency can be reduced whilst keeping the computational time below one minute.

The applications of wireless sensor networks, WSN, are getting popular in the different areas reaching from daily usage to industrial usage. The performance evaluation of WSN deployed in industrial and high-voltage areas is receiving a great attention and becoming an interesting area of research. This thesis addresses the performance issues of WSN in high-voltage and harsh industrial environments. This study has been carried out at the facilities of High Voltage Test Lab of ABB. Typically, wireless sensor network contains wireless field devices (nodes) connected to a base station via a central gateway. The gate way centralizes information gathered and processed by the nodes. The nodes can communicate with each other and with the gateway via radio wave. The quality and usability of the data sent by WSN can be degraded due to the packet loss and delay. In the presence of high-voltage, the electromagnetic interference, EMI, can affect the performance of WSN. In this study the performance of WSN is evaluated in terms of packet loss and delay. We also focus on the effect of EMI on hardware devices as well as on signal transmission. EMI was expected at wide frequency band due to harsh industrial and high voltage environments. It was expected that EMIs could increase a bit error rate and/or packet loss. The EMI can also change the sensitivity of the nodes. For the performance evaluation of WSN network throughput, latency, path stability, data reliability and average value of the received signal strength indicator, RSSI, are used and measured. The results show that the electromagnetic frequencies of harsh industrial and high voltage environments affect the wireless sensor network performance. / 0046700375216

Interactive 3D computer graphics requires a lot of computational resources to render a high quality frame. Typically the process of rendering a frame assumes a naïve approach that the whole frame can be perceived by the user in uniform detail. This is often not true, within 2° horizontal eccentricity from point of gaze is where one can primarily perceive details. Adjusting the quality of a frame based on the visual acuity can increase rendering performance by a factor of five to six at the resolution 1920x1080 without sacrificing perceived quality (Guenter et al., 2012a). Doing so without the user being aware of the manipulation requires a highly sophisticated system with low system latency able to update the display fast enough. The current study aims to answer what system latency is required to support passive gaze-contingent techniques that requires close to real-time gaze data. A unique experiment design was developed exposing test subjects to different system latencies by varying eye-tracker and monitor frequency. The outcome from the current study with 20 participants indicates a configuration with the estimated worst case system latency of 60ms is capable of hiding manipulation for 55% of the participants. Lowering the worst case system latency to 42ms and 95% of the participants reported that they could not detect any change. The study concludes that the configuration with estimated worst case system latency of 42ms is able to support passive gaze-contingent techniques. / Interaktiv 3D datorgrafik kräver mycket dataresurser för att producera en högkvalitativ bildruta. Den typiska processen att producera en bildruta antar det naiva antagandet att hela bildrutan kan uppfattas av användaren i enhetlig detalj. Detta antagande är ofta inte sant. Mänskliga ögat kan uppfatta detaljer i 2° horisontell vinkel från blickfokus med minskande detaljrikedom i ökande vinkel. Att anpassa kvaliteten på en bildruta baserat på detta kan öka renderingsprestanda med en faktor 5-6 vid upplösningen 1920x1080, utan att minska på den subjektivt upplevda kvaliteten (Guenter et al., 2012a). Att kunna göra det utan att användaren är medveten om manipulation kräver ett mycket sofistikerat system med låg system latens för att uppdatera displayen tillräckligt snabbt. Denna studie syftar till att svara på vad för system latens som krävs för att stödja tekniker som kräver nära realtid blick uppgifter. Studien presenterar ett unikt experiment där testpersoner exponeras för olika system latenser genom att variera både skärm och ögonspårnings frekvenser. Resultatet från studien som baseras på 20 deltagare visar att en konfiguration med estimerad värsta fall system latens på 60ms kan dölja manipulation för 55% av deltagarna. Vid 42ms system latens i värsta fall uppgav 95% av deltagarna att de inte kunde upptäcka någon förändring. Studien drar slutsatsen att konfigurationen med beräknade värsta fall systemet latens på 42ms kan stödja passiva blickberoende tekniker. / <p>Mobile: 073-2420226 Email: robin@thunstroem.com</p>

In this study we examine how much latency is required in an RTS-game (Real-Time Strategy) before the player quits the game. In the study latency is simulated in the games Age of Mythology and Starcraft: Brood War with the program WANem, and then the testplayers are asked “Do you want to quit because of latency?”. This data is then gathered and compiled into graphs to show the results. These results show that the limit for when players want to quit differs between the two games and that there are two factors that play into this. These factors are mainly the games mechanic, how the game is played, and how the game itself handles the latency. The results show that the limit for when players want to quit in Age of Mythology lies around 1000 ms, while in Starcraft: Brood War the limit is around 600 ms. / I denna studie undersöks det hur hög latency som krävs i ett RTS-spel (RealtidsStrategi) innan spelaren avslutar spelet. I studien simuleras latency i spelen Age of Mythology och Starcraft: Brood War med hjälp av programmet WANem, därefter får testpersonerna följande fråga ställda till sig: “Vill du sluta spela på grund av latency?”, denna data samlas sedan in och kompileras till grafer för att kunna visa resultaten. Resultaten som studien får fram är att gränsen skiljer sig mellan spelen och att det finns två faktorer som bidrar till detta, faktorerna är Spelets mekanik och Hur spelet hanterar latency. De siffror som studien har fått fram är att gränsen i Age of Mythology ligger på cirka 1000 ms medan gränsen i Starcraft: Brood War ligger på cirka 600 ms.

Artificial Neural Networks (ANNs) appear increasingly and routinely to gain popularity today, as they are being used in several diverse research fields and many different contexts, which may range from biological simulations and experiments on artificial neuronal models to machine learning models intended for industrial and engineering applications. One example is the recent success of Deep Learning architectures (e.g., Deep Belief Networks [DBN]), which appear in the spotlight of machine learning research, as they are capable of delivering state-of-the-art results in many domains. While the performance of such ANN architectures is greatly affected by their scale, their capacity for scalability both for training and during execution is limited by the increased power consumption and communication overheads, implicitly posing a limiting factor on their real-time performance. The on-going work on the design and construction of spike-based neuromorphic platforms offers an alternative for running large-scale neural networks, such as DBNs, with significantly lower power consumption and lower latencies, but has to overcome the hardware limitations and model specialisations imposed by these type of circuits. SpiNNaker is a novel massively parallel fully programmable and scalable architecture designed to enable real-time spiking neural network (SNN) simulations. These properties render SpiNNaker quite an attractive neuromorphic exploration platform for running large-scale ANNs, however, it is necessary to investigate thoroughly both its power requirements as well as its communication latencies. This research focusses on around two main aspects. First, it aims at characterising the power requirements and communication latencies of the SpiNNaker platform while running large-scale SNN simulations. The results of this investigation lead to the derivation of a power estimation model for the SpiNNaker system, a reduction of the overall power requirements and the characterisation of the intra- and inter-chip spike latencies. Then it focuses on a full characterisation of spiking DBNs, by developing a set of case studies in order to determine the impact of (a) the hardware bit precision; (b) the input noise; (c) weight variation; and (d) combinations of these on the classification performance of spiking DBNs for the problem of handwritten digit recognition. The results demonstrate that spiking DBNs can be realised on limited precision hardware platforms without drastic performance loss, and thus offer an excellent compromise between accuracy and low-power, low-latency execution. These studies intend to provide important guidelines for informing current and future efforts around developing custom large-scale digital and mixed-signal spiking neural network platforms.

The HIV-1 reservoir is the major barrier to eradication and cure of HIV-1 infection. The aim of this thesis was to characterise the HIV-1 reservoir in PHI and to investigate two potential reservoir clearance strategies. Sensitive Q-PCR assays that quantify HIV-1 DNA reservoir level were developed and applied to characterise the HIV-1 reservoir in acutely infected patients enrolled in the SPARTAC trial. HIV-1 DNA levels were found to be a useful biomarker for predicting clinical progression and the time of viral remission in patients, after HAART interruption. Since low HIV-1 DNA reservoir levels were clinically beneficial, potential therapies aimed at decreasing the reservoir burden were investigated. Thermotherapy was employed as a possible method for increasing CTL killing ability of HIV-1-infected cells. This may increase the likelihood of virological control and functional cure. SPION uptake impacted negatively on the natural killing ability of CTL that expressed the high affinity 868-TCR. However, localised thermotherapeutic heat generation was insufficient to cause direct thermal ablation of bound target cells or to enhance the natural killing ability of CTL. 868-TCR transduced CTL were employed directly as a means to target the latent viral reservoir by the ‘kick and kill’ hypothesis. Latent HIV-1 was reactivated in a variety of latent cell line models, but inconclusively in a primary model. 868-TCR transduced CTL killing of latently infected cells was observed. However, reactivation of latent virus did not necessarily relate to increased antigenicity of latent cells, perhaps in part due to anti-latency drug-induced alterations in MHC class I expression. Despite recent and sustained interest in HIV-1 cure strategies, the prospect still remains elusive. Only through the development of sensitive assays that measure the HIV-1 reservoir and their application to novel and innovative cure strategies will HIV-1 ever be functionally cured or eradicated in patients on a large scale.

In the field of communications, a high data rate and low multi-path fading is required for efficient information exchange. Orthogonal Frequency Division Multiplexing (OFDM) is a widely accepted IEEE 802.11n (and many others) standard for usage in communication systems operating in fading dispersive channels. In this thesis, we modeled the OFDM algorithm at the behavioral level in VHDL/Verilog that was successfully synthesized/verified on an FPGA. Due to rapid technology scaling, FPGAs have become popular and are low-cost and high performance alternatives to (semi-) custom ASICs. Further, due to reprogramming flexibility, FPGAs are useful in rapid prototyping. As per the IEEE standard, we implemented both transmitter and receiver with four modulation schemes (BPSK, QPSK, QAM16, and QAM64). We extensively verified the design in simulation as well as on Altera Stratix IV EP4SGX230KF40C2 FPGA (Terasic DE4 Development Board). The synthesized design ran at 100 MHz clock frequency incurring 54 µ sec. end-to-end latency and 8% logic utilization.

Nous abordons, dans cette thèse, les méthodes de combinaison de systèmesde transcription de la parole à Large Vocabulaire. Notre étude se concentre surl’attelage de systèmes de transcription hétérogènes dans l’objectif d’améliorerla qualité de la transcription à latence contrainte. Les systèmes statistiquessont affectés par les nombreuses variabilités qui caractérisent le signal dela parole. Un seul système n’est généralement pas capable de modéliserl’ensemble de ces variabilités. La combinaison de différents systèmes detranscription repose sur l’idée d’exploiter les points forts de chacun pourobtenir une transcription finale améliorée. Les méthodes de combinaisonproposées dans la littérature sont majoritairement appliquées a posteriori,dans une architecture de transcription multi-passes. Cela nécessite un tempsde latence considérable induit par le temps d’attente requis avant l’applicationde la combinaison.Récemment, une méthode de combinaison intégrée a été proposée. Cetteméthode est basée sur le paradigme de décodage guidé (DDA :Driven DecodingAlgorithm) qui permet de combiner différents systèmes durant le décodage. Laméthode consiste à intégrer des informations en provenance de plusieurs systèmes dits auxiliaires dans le processus de décodage d’un système dit primaire.Notre contribution dans le cadre de cette thèse porte sur un double aspect : d’une part, nous proposons une étude sur la robustesse de la combinaison par décodage guidé. Nous proposons ensuite, une amélioration efficacement généralisable basée sur le décodage guidé par sac de n-grammes,appelé BONG. D’autre part, nous proposons un cadre permettant l’attelagede plusieurs systèmes mono-passe pour la construction collaborative, à latenceréduite, de la sortie de l’hypothèse de reconnaissance finale. Nous présentonsdifférents modèles théoriques de l’architecture d’attelage et nous exposons unexemple d’implémentation en utilisant une architecture client/serveur distribuée. Après la définition de l’architecture de collaboration, nous nous focalisons sur les méthodes de combinaison adaptées à la transcription automatiqueà latence réduite. Nous proposons une adaptation de la combinaison BONGpermettant la collaboration, à latence réduite, de plusieurs systèmes mono-passe fonctionnant en parallèle. Nous présentons également, une adaptationde la combinaison ROVER applicable durant le processus de décodage via unprocessus d’alignement local suivi par un processus de vote basé sur la fréquence d’apparition des mots. Les deux méthodes de combinaison proposéespermettent la réduction de la latence de la combinaison de plusieurs systèmesmono-passe avec un gain significatif du WER. / This thesis presents work in the area of Large Vocabulary ContinuousSpeech Recognition (LVCSR) system combination. The thesis focuses onmethods for harnessing heterogeneous systems in order to increase theefficiency of speech recognizer with reduced latency.Automatic Speech Recognition (ASR) is affected by many variabilitiespresent in the speech signal, therefore single ASR systems are usually unableto deal with all these variabilities. Considering these limitations, combinationmethods are proposed as alternative strategies to improve recognitionaccuracy using multiple recognizers developed at different research siteswith different recognition strategies. System combination techniques areusually used within multi-passes ASR architecture. Outputs of two or moreASR systems are combined to estimate the most likely hypothesis amongconflicting word pairs or differing hypotheses for the same part of utterance.The contribution of this thesis is twofold. First, we study and analyze theintegrated driven decoding combination method which consists in guidingthe search algorithm of a primary ASR system by the one-best hypothesesof auxiliary systems. Thus we propose some improvements in order to makethe driven decoding more efficient and generalizable. The proposed methodis called BONG and consists in using Bag Of N-Gram auxiliary hypothesisfor the driven decoding.Second, we propose a new framework for low latency paralyzed single-passspeech recognizer harnessing. We study various theoretical harnessingmodels and we present an example of harnessing implementation basedon client/server distributed architecture. Afterwards, we suggest differentcombination methods adapted to the presented harnessing architecture:first we extend the BONG combination method for low latency paralyzedsingle-pass speech recognizer systems collaboration. Then we propose, anadaptation of the ROVER combination method to be performed during thedecoding process using a local vote procedure followed by voting based onword frequencies.

Organizational contexts use Situational judgment tests (SJTs) to assess and select individuals for competitive positions. As with other standardized assessments, threats to validity must be identified, examined, and communicated. As such, this research aims to identify the effects of socially desirable responding on the validity of an SJT used in a competitive selection process, and to identify if response latency provides insight into the identification of socially desirable responding. Participants in a competitive organizational selection process were administered an online assessment and the Balanced Inventory of Desirable Responding (BIDR). No significant correlations were identified between social desirability and the online assessment; however, this research provides evidence to suggest that participants who had shorter response latencies were less likely to participate in social desirability responding.